Douyin Video & Comment Scraper (抖音)

Get clean, structured data for any Douyin video from its URL or ID — or for the latest videos of a creator profile — the description, the creator, full engagement stats, the music, hashtags, the cover image and a direct video URL, plus optional comments and related videos. No Douyin account, no app, no manual copy-paste.

Douyin is China's TikTok (ByteDance), with ~750M daily users. Its data is hard to reach from outside the app and its API is request-signed. This actor signs requests itself in pure Python and, whenever Douyin's edge rejects that, retries the same call from inside a stealth browser that runs Douyin's own signer — so you get reliable structured JSON without paying for a browser on every request.

Who it's for

• Brand & agency teams — pull competitor videos, engagement, and hashtags.
• Creators & MCN analysts — benchmark a creator's latest videos against rivals.
• Researchers & trend analysts — study Chinese short-video content at scale.
• Data teams / AI apps — feed structured Douyin video data into pipelines.

Output fields

Field	Description
aweme_id	Video ID
desc	Caption / description
create_time	Publish time (unix seconds)
duration_ms	Video length in milliseconds
url	Canonical Douyin video URL
author	{ uid, sec_uid, unique_id, nickname }
stats	{ digg_count, comment_count, share_count, collect_count, play_count }
music	{ id, title, author }
hashtags	List of hashtag names
cover	Cover image URL
play_url	Direct video URL
width / height	Video dimensions
images	Image list for photo/note posts
comments	(optional) top-level comments: text, likes, reply count, user, location
relatedVideos	(optional) videos Douyin recommends alongside this one

Input

Field	Type	Notes
videoUrls	array	Video/note URLs or share links
videoIds	array	Bare numeric aweme IDs
profileUrls	array	Creator profiles (www.douyin.com/user/MS4wLjAB..., v.douyin.com/... share links, or bare sec_uid); each posted video becomes a row
maxVideosPerProfile	int	Cap videos per profile, newest first (default 20; Douyin shows logged-out visitors the ~40 most recent)
includeComments	bool	Attach top-level comments
maxCommentsPerVideo	int	Cap comments per video (default 50)
includeRelated	bool	Attach related videos
maxRelated	int	Cap related per video (default 10)
maxItems	int	Cap total videos
transport	enum	auto (default: Python first, browser fallback), python, browser
proxyConfiguration	proxy	Residential proxy recommended

Example input

{
  "videoUrls": [{ "url": "https://www.douyin.com/video/7682817863706627391" }],
  "profileUrls": [{ "url": "https://www.douyin.com/user/MS4wLjABAAAAOzRPcYBhm0xO0mXI5UGjHc2udHifSV-6xuI5mpo_UNKFN6qLVzfWb4QMT8BTOk2n" }],
  "maxVideosPerProfile": 20,
  "includeComments": true,
  "maxCommentsPerVideo": 20,
  "includeRelated": true,
  "proxyConfiguration": { "useApifyProxy": true, "apifyProxyGroups": ["RESIDENTIAL"] }
}

Example output (abridged)

{
  "aweme_id": "7682817863706627391",
  "desc": "特工夫妻深陷柏林死局… #柏林 #电影解说",
  "author": { "nickname": "心影说", "sec_uid": "MS4wLjABAA..." },
  "stats": { "digg_count": 8019, "comment_count": 79, "share_count": 634, "collect_count": 2720 },
  "duration_ms": 1041510,
  "hashtags": ["柏林", "电影解说"],
  "cover": "https://p3-pc-sign.douyinpic.com/...",
  "comments": [ { "text": "柏林谍变", "digg_count": 3, "ip_label": "天津" } ]
}

Scheduling & webhooks

Schedule the actor to re-scrape a set of videos or creator profiles and track engagement and new posts over time, or attach a webhook to push each run's rows into your database or Google Sheet.

Use with AI / MCP

Output is clean JSON — ideal as a tool for an AI agent analyzing Douyin content. Call the actor from the Apify MCP server or point your assistant at its dataset.

FAQ

Do I need a Douyin login? No. The actor scrapes public video data with its own anonymous, signed session.

Why a residential proxy? Douyin's edge is IP-sensitive; a residential proxy makes detail + comment fetching reliable. Datacenter IPs are often soft-blocked.

What does transport do? Leave it on auto. Requests are signed in pure Python (fast, cheap); if Douyin's edge rejects a request, the actor starts a stealth browser and repeats it there. python never starts a browser (cheapest, but intermittently blocked); browser uses the browser for everything (slowest, most robust). The actor also solves Douyin's proof-of-work interstitial itself.

Which URLs work? www.douyin.com/video/<id>, /note/<id>, ?modal_id=<id>, and share links (v.douyin.com/..., iesdouyin.com/share/video/...) in videoUrls; www.douyin.com/user/<sec_uid> and profile share links in profileUrls.

How many videos per creator profile? Up to about the 40 most recent posts. That is all Douyin shows a logged-out visitor; older posts need a Douyin login, which this actor does not use. Re-run on a schedule to keep collecting new posts.

Keyword search? Douyin's web search is login-walled, so this actor works by video URL/ID or creator profile. Use the Douyin Hot Search Tracker for trending terms.

Is this legal? The actor reads publicly visible video data. You are responsible for using the output in line with Douyin's terms and your local laws.